Temporal and spatial variations of precipitation in Northwest China during 1960–2013

• The precipitation in northwest China had a significantly increasing trend (P < 0.01) for annual rainfall, at a rate of 0.55 mm/a.
• It is not necessarily that while the periodic components have significant trends, the original data also have.
• The drivers from periodic component of original data was different from season to season, and different from area to area.
• Most of the trends, whether positive or negative, started during the mid − 1980s to mid − 1990s.
• The relationships between the precipitation and SST are different for different periodic components and different periods.

Based on the precipitation data from 96 weather stations in Northwest China (NWC) during 1960–2013, the Continuous Wavelet Transform (CWT) and the Mann-Kendall (MK) test were applied to analyze the precipitation spatiotemporal variations at different time scales. The relationships between the original precipitation and different periodic components were investigated. The results indicated that the annual precipitation was significantly increasing (P < 0.01) at the rate of 0.55 mm/a in the NWC. In terms of seasonal precipitation, the summer original precipitation significantly increased (P < 0.05) in the Southern Altay Mountain Basin (SAMB), Qaidam Basin (QB), Qiang Tang Plateau Basin (QTPB), Turpan-Hami Basin (THB), Tarim Desert Basin (TDB), Northern Tianshan Mountain Basin (NTMB) and NWC. For the winter original precipitation, except the Inner Mongolia Inland Rivers Basin and Northern Kunlun Mountain Basin, the significant increases (P < 0.05) were detected in the other sub-basins. In terms of monthly precipitation, significant increases were detected in January in the SAMB, NTMB and NWC, and July in the QB, Headstreams of Tarim River Basin (HTRB) and N. Additionally, most of the increasing and decreasing trends began in the mid-1980s or mid-1990s. Moreover, the periodic components were not always similar to the original data with the significant trends. The dominant scale of the original data from the periodic components was different in spatiotemporal distribution. Meanwhile, the relationship between the precipitation and El Niño-Southern Oscillation (ENSO) was different from period to period and from time scale to time scale. This study will help to develop better management measures to account for climate change and the supply/demand of water.